search

FredHutch / SEACR

SEACR: Sparse Enrichment Analysis for CUT&RUN
GNU General Public License v2.0
101 stars 45 forks source link

Getting two AUC.BED files after bash SEACR #42

Open ahua217 opened 4 years ago

ahua217 commented 4 years ago

Hello! My question is about the difference of webpage use of SEACR and bash SEACR in terminal. They gave out different output files. My setting is "non" and "stringent". From webpage, I got only one .BED file, which I think showed the peaks of target subtracted with that of IgG. Whereas from bash SEACR, I got two .AUC files and two .AUC.BED files, which I did not know how to annotate. I used "bash SEACR_1.3.sh target.bedgraph IgG.bedgraph non stringent output". Should I do some changes in bash use?

Thank you very much!

mpmeers commented 4 years ago

Hi,

The .AUC and .AUC.bed files are intermediate files that should be erased before the command finishes, meaning that an error of some sort likely cropped up. If the command completes as expected, you should only get one .bed file output, which in your case would be called "output.stringent.bed". If you run the command again, is there an error message that appears in the standard output?

Mike

ahua217 commented 4 years ago

Thank you for the help! I run it again and this time I copied what returned

BED2BGP ahua217$ bash SEACR_1.3.sh 137665_2.bedgraph 137668_2.bedgraph non stringent output Calling enriched regions with control file Proceeding without normalization of control to experimental bedgraph Using stringent threshold Creating experimental AUC file: Sat Sep 19 14:43:10 EDT 2020 Creating control AUC file: Sat Sep 19 14:44:36 EDT 2020 Calculating optimal AUC threshold: Sat Sep 19 14:46:17 EDT 2020 Calculating threshold using non-normalized control: Sat Sep 19 14:46:17 EDT 2020 dyld: Library not loaded: @rpath/libreadline.6.2.dylib Referenced from: /Users/ahua217/miniconda3/lib/R/lib/libR.dylib Reason: image not found /Users/ahua217/miniconda3/bin/SEACR_1.3.sh: line 130: 11588 Abort trap: 6 Rscript $path/SEACR_1.3.R --exp=$password.auc --ctrl=$password2.auc --norm=no --output=$password

It looks the problem is library not loaded and image not found. Did I miss any setting previously? I am fresh about it.

ahua217 commented 4 years ago

Though webpage of SEACR works for me, I am trying to realize batch analysis if I know how to use bash.

mpmeers commented 4 years ago

Hi,

Do you have R in your path? It seems like it's tripping up when the R script is executed. There are not special libraries within R that you should need, so I suspect there's an issue with R itself.

Mike

ahua217 commented 4 years ago

Today I got a chance to turn off Rstudio but the issue was still there. Then I found a similar post leading to a potential solution

conda update -c rdonnellyr -c main --all

I have not tried it because I am not sure if this command affects other software and my batch analysis with SEACR is not urgent now.